detection depth

[Ferric Toes]

Put both charts(fixed inductance) on same page. Both give the same results, detection distance vs coil size. Both assume everything remains the same but coil radius(diameter) and noise level doesn't change. Don't need math to chart the left chart. Been trying to understand the best way to use the left chart. With the Excel chart I can read change in signal strength or detection distance when changing coil size. Haven't figured how to read change in signal strength with the left chart.

Here is a page from my 1978 notebook. It's late here now, so I will do an explanation tomorrow, although I expect Skippy can figure it out.

Eric.

[Skippy]

I'm not entirely sure what info Green is wanting, it appears like he wants Target Signal strength plotted against coil radius, for a family of 'target sizes' , that match the 'target sizes' used on the semi-circle chart. So they would look vaguely like a series of inverted parabolas, peaking at the 'max depth' point. Such a chart could be plotted with the same 'coil radius' scale, so the two would line up. Would that be useful?

[green]

I'm not entirely sure what info Green is wanting, it appears like he wants Target Signal strength plotted against coil radius, for a family of 'target sizes' , that match the 'target sizes' used on the semi-circle chart. So they would look vaguely like a series of inverted parabolas, peaking at the 'max depth' point. Such a chart could be plotted with the same 'coil radius' scale, so the two would line up. Would that be useful?

A chart like I think I want. Charted with a coil diameter increase of 1.5 times starting with 133mm coil. 133, 200, 300, 450 and 675mm coils all same inductance. I like the Excel chart because I can read signal change or distance change with coil diameter change. Can estimate needed increase in Tx peak current to double detection distance or needed increase in Tx peak current to increase detection distance by(maybe 50mm). The other chart is easy to make but I have a harder time reading it, maybe with more practice. Still don't know how to make it so I could read change in signal strength and change in detection distance.
Maybe just me but the first thing I have to do is convert radius to diameter when I think about it. Maybe radius is better for others.

[Ferric Toes]

A chart like I think I want. Charted with a coil diameter increase of 1.5 times starting with 133mm coil. 133, 200, 300, 450 and 675mm coils all same inductance. I like the Excel chart because I can read signal change or distance change with coil diameter change. Can estimate needed increase in Tx peak current to double detection distance or needed increase in Tx peak current to increase detection distance by(maybe 50mm). The other chart is easy to make but I have a harder time reading it, maybe with more practice. Still don't know how to make it so I could read change in signal strength and change in detection distance.
Maybe just me but the first thing I have to do is convert radius to diameter when I think about it. Maybe radius is better for others.

Geometrically, the target is on the centre line to give maximum signal, so that is why radius is used. Detectorists normally talk in terms of coil diameter though, so then you have to remember to halve it. In my old graph that Skippy posted a while back, I used 'coil diameter' on the x axis and find it more difficult to read and remembering to halve it.

I have only ever used that chart, or the associated calculations, a few times when designing for specific targets. There was a case when a detectorist found a scattered hoard of 13th century silver coins on the beach. Because coins go deep in beach sand and a few had been found, we could test one coin on a certain size coil and then work out whether a larger coil would be more effective. For general searching for a wide spectrum of object sizes and conductivities, my view is that the graph has limited use.

The final detection range is also modified by several variables i.e. number of turns (inductance), coil current, sample delay, sample width, and filter characteristics, all of which can be optimised for any particular single object.

Eric.

[green]

Geometrically, the target is on the centre line to give maximum signal, so that is why radius is used. Detectorists normally talk in terms of coil diameter though, so then you have to remember to halve it. In my old graph that Skippy posted a while back, I used 'coil diameter' on the x axis and find it more difficult to read and remembering to halve it.

I have only ever used that chart, or the associated calculations, a few times when designing for specific targets. There was a case when a detectorist found a scattered hoard of 13th century silver coins on the beach. Because coins go deep in beach sand and a few had been found, we could test one coin on a certain size coil and then work out whether a larger coil would be more effective. For general searching for a wide spectrum of object sizes and conductivities, my view is that the graph has limited use.

The final detection range is also modified by several variables i.e. number of turns (inductance), coil current, sample delay, sample width, and filter characteristics, all of which can be optimised for any particular single object.

Eric.

I have a question on amplifier gain. Coil inductance, peak current, coil diameter, amplifier gain and others effect if a target would saturate the amplifier. My detector, 300uH coil, 1amp peak current, 200mm diameter, 450 amplifier gain.
A 20mm square(1/10 the diameter) cut from an aluminum coke can laying on the coil center just saturates the amplifier. If I increased the peak current to 5amps should I reduce the gain to 1/5th? Is there a test to determine if amplifier gain is near what it should be? Maybe a square cut from an aluminum can 1/10 the coil diameter laying on the coil just saturates the amplifier as an example.

[Ferric Toes]

I have a question on amplifier gain. Coil inductance, peak current, coil diameter, amplifier gain and others effect if a target would saturate the amplifier. My detector, 300uH coil, 1amp peak current, 200mm diameter, 450 amplifier gain.
A 20mm square(1/10 the diameter) cut from an aluminum coke can laying on the coil center just saturates the amplifier. If I increased the peak current to 5amps should I reduce the gain to 1/5th? Is there a test to determine if amplifier gain is near what it should be? Maybe a square cut from an aluminum can 1/10 the coil diameter laying on the coil just saturates the amplifier as an example.

Hi Green,

Some designs use diodes in the feedback loop that bring in lower value resistors when the diode forward voltage is exceeded. I have tried that, but not found any real benefit. The venerable NE5534 comes out of saturation fast enough to have no effect if used as a single amplifier for 450x gain. Faster amplifiers will give a bit more leeway on delays of <10uS Mostly it is the d.c. amplifier after the integrator that saturates on strong signals, but again there is no ill effect. Often a gain control is on the d.c. amplifier.

Increasing the TX current from 1A to 5A will increase the detection range by 1.3 times whatever it was before. What comes out of all this is to measure the range that you can just detect your 20mm square with your 100mm radius coil and if it is considerably more than 100mm, say it was 200mm, then using the graph the curves would indicate that increasing the coil radius to 200mm would gain another 45mm in range at a coil current of 1A. Keeping the inductance the same and having 5A current would give a total increase of 1.3 x 245mm. Your new range would now be 318mm with a 400mm diameter coil.

Eric

[green]

Hi Green,

Some designs use diodes in the feedback loop that bring in lower value resistors when the diode forward voltage is exceeded. I have tried that, but not found any real benefit. The venerable NE5534 comes out of saturation fast enough to have no effect if used as a single amplifier for 450x gain. Faster amplifiers will give a bit more leeway on delays of <10uS Mostly it is the d.c. amplifier after the integrator that saturates on strong signals, but again there is no ill effect. Often a gain control is on the d.c. amplifier.

Increasing the TX current from 1A to 5A will increase the detection range by 1.3 times whatever it was before. What comes out of all this is to measure the range that you can just detect your 20mm square with your 100mm radius coil and if it is considerably more than 100mm, say it was 200mm, then using the graph the curves would indicate that increasing the coil radius to 200mm would gain another 45mm in range at a coil current of 1A. Keeping the inductance the same and having 5A current would give a total increase of 1.3 x 245mm. Your new range would now be 318mm with a 400mm diameter coil.

Eric

Some of the build detectors have an amplifier gain of 1000. Carl replied in another thread that he used a gain of 100 I think. Haven't been able to find the thread. There has to be some logic in selecting amplifier gain. Since coil current and others including amplifier gain effect signal output I was wondering what the thinking process is or if there was a test to determine if the total gain is in the ball park.

[Skippy]

If you're going to be using a variety of coils, and are looking for a variety of targets, it seems almost a necessity to have some gain control.
This is where the difference between a monocoil and say a DD, show up. The DD shows a major increase in signal strength as the target gets towards Z = 0, whereas the mono is much less affected. An overloaded DD can simply be raised a little, and the signal drops away. The mono needs to be raised further, though the technique still works, of course.
I've only got VLF experience, but my feeling is that a small square of drinks-can shouldn't be on the 'overload threshold' when in the centre of a coil. Maybe at Z=5cm distance would be more appropriate. You're typically going to have the coil 2.5cm above the dirt/soil anyway, I guess.

[Ferric Toes]

Some of the build detectors have an amplifier gain of 1000. Carl replied in another thread that he used a gain of 100 I think. Haven't been able to find the thread. There has to be some logic in selecting amplifier gain. Since coil current and others including amplifier gain effect signal output I was wondering what the thinking process is or if there was a test to determine if the total gain is in the ball park.

I used to have a gain of 1000 for the preamp in the days when 50uS was considered a short delay and coils were unshielded. As delay times were reduced to detect smaller and less conductive objects, front end gain was reduced to around 500. This was because the opamp then had a wider bandwidth which resulted in a faster recovery time from the TX transients.

If you look at opamp data sheets, there is usually a graph showing open loop gain and bandwidth. Reducing gain from 60db (1000x) to 40db (100x) results in more flat bandwidth for the preamp at the lower gain. This loss of gain can be made up in the later stages i.e. the d.c. amplifier. Recovery speed of the front end opamp is a combination of small signal bandwidth and high signal slew rate.

Today though, there is a much greater choice of front end amplifiers and it is not difficult to find one that will give a gain of 1000 with a more than adequate bandwidth of several MHz. However, there is no free lunch and such devices are often noisy at low frequencies and current hungry.

Modern detector designs sometimes use a gate arrangement that disconnects the preamp from the coil until the TX transients have finished. The Vallon VMH3CS appears to have this feature.

Eric.

[green]

Did a chart with stacked quarters awhile back. Another with quarters side be side, not stacked. Three small divisions doubles signal. 2coins twice signal 1coin. 4coins twice signal 2coins. Number of coins doesn't effect decay TC.

[green]

Charted clad and silver dimes and quarters. Quarters chart closer than dimes. Wondering if a VLF detector shows a greater difference with clad and silver dimes. Quarters almost the same.

[Skippy]

I guess this question is aimed at me. And the answer is yes, clad dimes read different to 90% silver ones. Disclaimer: I only have ONE silver dime, a lightly circulated 'Mercury' one. My guess is the 10c cladding thickness is the same as on the 25c coin, hence it is proportionally more significant compared to what it is on the 25c. Ie. less good-conductor copper core, more poor conductor CuNi skin.
I have measurements as:
Mercury 10c : f_c = 2.5 kHz; tc = 64 usec
Clad 10c : f_c = 2.75 kHz; tc = 58 usec

As this is a thin coin, my 13kHz VLF figures may match your PI ones better, due to skin depth issues

[green]

I guess this question is aimed at me. And the answer is yes, clad dimes read different to 90% silver ones. Disclaimer: I only have ONE silver dime, a lightly circulated 'Mercury' one. My guess is the 10c cladding thickness is the same as on the 25c coin, hence it is proportionally more significant compared to what it is on the 25c. Ie. less good-conductor copper core, more poor conductor CuNi skin.
I have measurements as:
Mercury 10c : f_c = 2.5 kHz; tc = 64 usec
Clad 10c : f_c = 2.75 kHz; tc = 58 usec

As this is a thin coin, my 13kHz VLF figures may match your PI ones better, due to skin depth issues

Thanks, I checked the slope on the chart and got 76usec and 85usec for the dimes. Have you checked a clad and silver quarter to see I they test closer?

[Skippy]

At least we agree they are different, by about 10%, even if, as usual, the actual figures don't match.

And you should be able to find previous discussion about the 25c coin somewhere - they are pretty much identical to a VLF, though again, my tc figures didn't match your PI ones..

[green]

At least we agree they are different, by about 10%, even if, as usual, the actual figures don't match.

And you should be able to find previous discussion about the 25c coin somewhere - they are pretty much identical to a VLF, though again, my tc figures didn't match your PI ones..

I did find your reply, thanks